Artificial Neural Network for Misuse Detection
Laheeb Mohammed Ibrahim · Journal of Communication and Computer · 2010
With the growing numbers of attackers on computer networks, detecting new attacks is not an easy job for intrusion detection system (IDS), an intelligent method like artificial neural networks (ANN) has been a successful method to solve this problem. In this paper, we create a misuse IDS system to show that unsupervised growing hierarchical self-organizing maps (GHSOM) and recurrent Jordan-Elman neural networks perform surprisingly well when aggressively trained and tested on a corpus of attacks to detect and classify known and new attacks in network traffic.